|
|
|
|
|
|
Re: What new retrobrew projects are people interested in? [message #2960 is a reply to message #2959] |
Mon, 05 June 2017 06:20 |
|
lowen
Messages: 226 Registered: August 2016 Location: Western NC USA
|
Senior Member |
|
|
Well, my own experience with reviving the CPU280 project tells me that people sometimes are interested in what was 'unobtainium' during their own early years in computers. I know that's why I've gone through the process of reviving the CPU280 build, once I found sources for all the parts. 'Legendary' CPUs such as the Z280, where there was so much promise but such an underwhelming delivery are interesting; but, then again, I've been excited about the Z280 since it was known as the Z800 and the TRS-80 Model 4 non-gate-array board had extra holes and logic to use the 8-bit bus version of the Z800, 32+ years ago. TMS9900 is another such chip, where the promise of 16-bits was severely hampered by the implementation.
Personally, if I had more disposable income and time, I would probably try John's KISS-68030 as a next step. The Z280 is equivalent in power to an 8086 for most things, and it's related to the Z80. I have more x86 machines than I can shake a stick at already, and the only reason I can think of to build an x86 is to compare its performance to the CPU280, but I can do that with the NEC V20 in my Tandy 1400LT. 68K, on the other hand, is interesting, and I might even think about building an SBC workalike to the Tandy 6000, but with the hardware pieces I want to add, and then hack Xenix to work on it.....
If the chips were available, an iAPX432 might be neat. A System/370 would be really cool.
Barring that, a DEC J11 or F11 hobbyist-buildable SBC might be fun, but using a bus that people already have, either ISA or ECB or maybe S-100. While I think Multibus is a great architecture, at some point having too many diverging projects that aren't able to leverage existing investment in peripheral boards, backplanes, and enclosures is going to dilute interest. One of the great things about the original 'N8VEM' Z80 board was the use of a semi-standard bus that isn't terribly expensive to implement on a hobbyist board.
ECB, even with its quirks and nonstandard extensions, was an excellent choice, but it is more excellent for a Z80 than for other chips (even the CPU280, which runs the Z280 chip, needs special logic to interface to the Z80-ish ECB due to the design using the 16-bit Z-BUS mode of the Z280; but Tilmann's solution to getting an I/O-port only fully Mode2-compatible ECB from a Z-BUS Z280 is neat).
At some point, there is a delicate balance between chip availability and the ability for the hobbyist to work with the chips. We are at the inflection point with the Z280, where apparently a whole lot of the chips have become available, but I wouldn't count on long-term availability. The eZ80 is very available and would be a neat chip for a very capable SBC, but the packaging of that chip presents significant challenges to the hobbyist. PLCC-84 is the largest through-hole capable SMT package I know of, and that is a significant limitation. EDIT: PGA is normally through-hole, but once you go to the staggered PGA of the Pentium trace routing becomes devilishly difficult at the socket with anything less than 4 or 6 layers.
At the moment, the Z80 family is my main interest; I already know how to program the Z80, and I don't have time at the moment to learn a whole new architecture.
IMHO, YMMV, etc.
--
Bughlt: Sckmud
Shut her down Scotty, she's sucking mud again!
[Updated on: Mon, 05 June 2017 06:45] Report message to a moderator
|
|
|
Re: What new retrobrew projects are people interested in? [message #2961 is a reply to message #2958] |
Mon, 05 June 2017 07:32 |
jcoffman
Messages: 332 Registered: October 2015
|
Senior Member |
|
|
David,
I build an S-100 system using the NS32016 w/ TCU, MMU, FPU, ICU. This was many years ago, and the size of SRAM chips at the time was 32K x 8 max. Later, I purchased a used OPUS board with the NS32532 & NS32382. I did a little playing with both boards, including an assembler which first ran on an Apple II, later on a PC/AT. Sometime later, while fooling around with LCC 3.1, I had a full-blown C-compiler running. The limited memory of the old S-100 system kept me from ever doing very much with the boards.
Your link to the info pages on all the chips has proved most helpful.
--John
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Re: What new retrobrew projects are people interested in? [message #2983 is a reply to message #2982] |
Tue, 06 June 2017 11:43 |
lynchaj
Messages: 1051 Registered: June 2016
|
Senior Member |
|
|
The PC532 is a neat design but will only appeal to the most dedicated and wealthy hobbyists. It is a large 6 layer PCB with many expensive chips (NS32532, NS support chips, 2 different SCSI hosts, etc.)
Is there anything left of the old PC532 mailing list/community? There may be others out there who'd like an opportunity for renewed hardware development.
What would be ideal is a less expensive design using a smaller 2 or 4 layer PCB, CPU with integrated support chips, CPLD or FPGA for glue logic, etc. while retaining software compatibility.
[Updated on: Tue, 06 June 2017 12:17] Report message to a moderator
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Re: What new retrobrew projects are people interested in? [message #3002 is a reply to message #3001] |
Wed, 07 June 2017 04:42 |
lynchaj
Messages: 1051 Registered: June 2016
|
Senior Member |
|
|
Another possibility for an OS, although rather unlikely, would be to port the CP/M-68K C sources over to NS32CG160. I believe the CP/M-68K sources have been ported to other architectures before for this purpose. It would not be a simple task for sure but it might be easier than wrangling with NetBSD. Adding a serial port, 1-2 MB of SRAM, and an IDE interface to the NS32CG160 design in the application notes may be sufficient to get a simple CP/M-NS32K system running. Obviously it would require some kind of NS32K monitor to get started though.
https://groups.google.com/forum/#!topic/comp.os.cpm/RUIgiyNy LmI
I am willing to help out on any NS32K design if it comes to pass.
[Updated on: Wed, 07 June 2017 04:50] Report message to a moderator
|
|
|
|
Re: What new retrobrew projects are people interested in? [message #3004 is a reply to message #3003] |
Wed, 07 June 2017 07:06 |
|
lowen
Messages: 226 Registered: August 2016 Location: Western NC USA
|
Senior Member |
|
|
RE: NS32K OS:
NetBSD is a known system and extremely well documented. Whole books have been written on its design and implementation (the 'Daemon Book', aka 'The Design and Implementation of the 4.3BSD UNIX Operating System,' is one such book, and a pretty clear read (yes, I have a copy, but I'm not going to pretend to understand it all!)).
This might be a porting target for Fuzix.
CP/M-68K ported might also be interesting as a tool to learn the chip.
Andrew L, the gerbers are in the file et532-pcb.tar.gz in the FTP directory on nic.funet.fi; you'll want to browse around that whole tree! EDIT/CORRECTION: As pointed out downthread, the gerbers here are for the Ethernet card; my error.
--
Bughlt: Sckmud
Shut her down Scotty, she's sucking mud again!
[Updated on: Thu, 08 June 2017 09:26] Report message to a moderator
|
|
|
Re: What new retrobrew projects are people interested in? [message #3005 is a reply to message #3002] |
Wed, 07 June 2017 07:20 |
jcoffman
Messages: 332 Registered: October 2015
|
Senior Member |
|
|
lynchaj wrote on Wed, 07 June 2017 04:42Another possibility for an OS, although rather unlikely, would be to port the CP/M-68K C sources over to NS32CG160. I believe the CP/M-68K sources have been ported to other architectures before for this purpose. It would not be a simple task for sure but it might be easier than wrangling with NetBSD. Adding a serial port, 1-2 MB of SRAM, and an IDE interface to the NS32CG160 design in the application notes may be sufficient to get a simple CP/M-NS32K system running. Obviously it would require some kind of NS32K monitor to get started though.
https://groups.google.com/forum/#!topic/comp.os.cpm/RUIgiyNy LmI
I am willing to help out on any NS32K design if it comes to pass.
CP/M-68 is largely written in 'C' for the big-endian 68000; the NS32000 series is little endian, like Intel and Zilog. CP/M-68 comes with good documentation on the BIOS calls needed to run it. This is a major plus.
CP/M-68 was first brought up on the Mini-M68k, and I can testify that it was a relatively easy port from the C-dialect in which it is written to the GCC/68000 cross-compiler. That was several years ago, and I must say I don't really remember how much attention the source pays to "endian-ness". It must pay quite a bit, because the file-system it implements is compatible with Zilog CP/M up to their 8mb limit. CP/M-68 supports much larger partitions, to 512mb, if I remember correctly.
Has GCC ever supported the NS32000 opcodes? I truly do not know.
--John
|
|
|
|
|